RESEARCH ARTICLE Open Access

Antibiotics resistance and toxin profiles ofBacillus cereus-group isolates from freshvegetables from German retail marketsGregor Fiedler1* , Carmen Schneider2, Etinosa O. Igbinosa1,3, Jan Kabisch1, Erik Brinks1, Biserka Becker2,Dominic A. Stoll2, Gyu-Sung Cho1, Melanie Huch2 and Charles M. A. P. Franz1

Abstract

Background: This study aimed to evaluate the safety of raw vegetable products present on the German marketregarding toxin-producing Bacillus cereus sensu lato (s.l.) group bacteria.

Results: A total of 147 B. cereus s.l. group strains isolated from cucumbers, carrots, herbs, salad leaves and ready-to-eat mixed salad leaves were analyzed. Their toxinogenic potential was assessed by multiplex PCR targeting thehemolysin BL (hbl) component D (hblD), non-hemolytic enterotoxin (nhe) component A (nheA), cytotoxin K-2 (cytK-2) and the cereulide (ces) toxin genes. In addition, a serological test was used to detect Hbl and Nhe toxins. On thebasis of PCR and serological results, none of the strains were positive for the cereulide protein/genes, while 91.2,83.0 and 37.4% were positive for the Hbl, Nhe and CytK toxins or their genes, respectively. Numerous strainsproduced multiple toxins. Generally, strains showed resistance against the β-lactam antibiotics such as penicillin Gand cefotaxim (100%), as well as amoxicillin/clavulanic acid combination and ampicillin (99.3%). Most strains weresusceptible to ciprofloxacin (99.3%), chloramphenicol (98.6%), amikacin (98.0%), imipenem (93.9%), erythromycin(91.8%), gentamicin (88.4%), tetracycline (76.2%) and trimethoprim/sulfamethoxazole combination (52.4%). Thegenomes of eight selected strains were sequenced. The toxin gene profiles detected by PCR and serological testmostly agreed with those from whole-genome sequence data.

Conclusions: Our study showed that B. cereus s.l. strains encoding toxin genes occur in products sold on the Germanmarket and that these may pose a health risk to the consumer if present at elevated levels. Furthermore, a smallpercentage of these strains harbor antibiotic resistance genes. The presence of these bacteria in fresh produce should,therefore, be monitored to guarantee their safety.

Keywords: Bacillus cereus sensu lato, Fresh produce, Toxins, Antibiotic resistance, Whole genome sequencing, Foodsafety

BackgroundThe Bacillus (B.) cereus group is of great importance tothe food industry as it is associated with food spoilage,resulting in deterioration of food quality, as well as offood safety, affecting public health and the economy.Bacterial toxins ranked second in 2016 as a group ofcausative agents for foodborne outbreaks in Europe [1].Bacillus cereus group species are readily isolated from

food crops and food plants [2–4]. Significant numbers ofbacteria belonging to the B. cereus group have beenshown to occur in dried spices and herbs, but also infresh herbs and vegetables (carrots, leaf lettuce, cucum-bers and mixed salad leaves) in Germany [2, 5, 6]. Thepresence of both vegetative cells and spores in foodcommodities has been reported and their role in foodsafety and food spoilage is well known [7]. In 2016, theEuropean Union (EU) reported 413 foodborne outbreaksbeing caused by Bacillus toxins affecting 6657 people.Of these 352 people were hospitalized without terminaloutcome [7]. However, certain severe cases of foodborne

© The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

* Correspondence: gregor.fiedler@mri.bund.de1Department of Microbiology and Biotechnology,Hermann-Weigmann-Straße 1, 24103 Kiel, GermanyFull list of author information is available at the end of the article

Fiedler et al. BMC Microbiology (2019) 19:250 https://doi.org/10.1186/s12866-019-1632-2

illness by B. cereus leading to mortality or organ fail-ure have sporadically been reported in several coun-tries [8–10].Bacillus cereus belongs to the B. cereus group of bac-

teria, also known as B. cereus sensu lato (s.l.), which con-sists of eight genetically closely related species thatinclude B. anthracis, B. cereus sensu stricto (s.s.), B. cyto-toxicus, B. mycoides, B. pseudomycoides, B. thuringiensis,B. toyonensis and B. weihenstephanensis [11–14]. Thetaxonomic relationships between members of the B. ce-reus group are controversially debated and identificationbased on 16S rRNA gene sequencing fails due to highnucleotide conservation in the sequence of this geneamong species of this group [14]. The species occurringwithin the B. cereus s.l. have so far been distinguished onthe basis of morphological, physiological and/or viru-lence characteristics, the latter which includescharacterization of toxin genes that are often located onextrachromosomal genetic elements (plasmids), as wellas multilocus sequence analysis and phylogenomics [14].Bacillus cereus s.s. is an opportunistic pathogen cap-

able of causing a range of diseases [15], most promin-ently foodborne disease due to the production ofenterotoxins (diarrheal toxin) or a non-ribosomal pep-tide synthetase (NRPS) toxin (emetic or cereulide toxin)[16]. The diarrheal toxins hemolysin BL (Hbl) non-hemolytic enterotoxin (Nhe) and cytotoxin K (CytK) [17,18] have been linked to the diarrheal type of B. cereusfood poisoning, which is characterized by abdominalpain and watery diarrhea. The Hbl and Nhe enterotoxinsare three-component toxins consisting of the subunitsHblC, HblD and HblA for Hbl toxin and the subunitsNheA, NheB and NheC for the Nhe toxin. The virulenceof emetic strains is attributed to the production of aheat-stable cereulide, synthesized by a non-ribosomalpeptide synthetase encoded by ces genes [16].The growing challenge of the emergence and spread of

antibiotic resistance has been the subject of several sur-veys [19] and has been reported by the World HealthOrganizations as one of the major health challenges ofthe twenty-first century [20, 21]. Bacillus cereus s.s. istypically resistant to penicillin and other β-lactam antibi-otics [22] and can furthermore acquire resistance tocommonly used antibiotics such as ciprofloxacin, cloxa-cillin, erythromycin, tetracycline and streptomycin [22,23]. Foodborne illness associated with B. cereus groupstrains seldomly need to be treated with antibiotics.However, it has not been well investigated to which de-gree B. cereus group-strains can serve as a source oftransferable antibiotic resistance genes in the food chain.In this study, bacterial strains from fresh vegetables on

the German market were identified as belonging to theB. cereus s.l. group by 16S rRNA gene sequencing, andtheir potential for toxin production was determined.

This was done to determine the toxinogenic potential ofthese organisms in fresh produce. For this, a multiplexPCR based on the toxin genes hblDA, nheAB, cytk andces as target genes was used. The multiplex PCR wasdone in independent replicates in two different labora-tories with different polymerase enzymes to determinethe reproducibility of the methods. Apart from molecu-lar analysis, the expressed bacterial toxins NheB andHblC (HblL2) were also detected using a commerciallyavailable immunological assay. In addition, the pheno-typic antibiotic resistances of strains were determinedand the antibiotic resistance genes of selected resistantisolates were characterized by whole genome sequen-cing. Furthermore, whole genome sequence data wereused to investigate differences in toxin gene typing by ei-ther PCR or serological testing. Therefore, the studyaimed to provide an indication on the risk of B. cereus s.l.strains in vegetables on the German market from both afood safety and antibiotic resistance point of view.

ResultsPhenotypic and genotypic characterizationB. cereus-group isolates were characterized according totheir typical colony characteristics and to their growthbehavior at different temperatures. Only very few (7 of147 strains, 4.8%) were able to grow at 4 °C, while amoderate percentage of strains were able to grow at 7 °C(45 out of 147, 30.6%) and at 50 °C (40 out of 147,27.2%). Nearly all strains showed β-hemolysis, except 2out of 147 strains (1.4%) which were not hemolytic (γ-hemolytic). The 16S rRNA genes of all strains were se-quenced and the strains all clustered closely togetherwith the type strains of the B. cereus s.l. group, namelyB. cereus s.s., B. toyonensis, B. weihenstephanensis, B.anthracis, B. mycoides, pseudomycoides and B. thurin-giensis at r = 98.8% (Additional file 1: Figure S1). Thetype strains of B. cytotoxicus clustered at r = 97.6% withthe above mentioned strains. On the other hand, the 16SrRNA gene sequences clustered apart from the typestrain of B. subtilis, showing that all isolates in this studyshould be considered as belonging to the B. cereus s.l.group.Eight strains of the B. cereus s.l. group isolates were se-

lected as they either showed resistance towards one ormore antibiotics, or because they represented particularenterotoxin types. The genome sequences of the eightselected strains were also used for genome-based phylo-genetic comparison and the strains were clustered intoone of three groups (Fig. 1). Four strains (532a, MS12,MS464a and G12) clustered together with the B. cereusATCC 14579T (DSM 31T) and the B. thuringiensisATCC 10792T (DSM 2046T) type strains. Two strains(MS735 and MS195) clustered together with the B. toyo-nensis BCT-7112T type strain, while two (MS17 and

Fiedler et al. BMC Microbiology (2019) 19:250 Page 2 of 13

B26) clustered with the B. mycoides ATCC 6462T (DSM2048T) and the B. weihenstephanensis WSBC 10204T

(DSM 11821T) type strains (Fig. 1).

Toxin gene profilesThe toxin gene profiles established in this study usingmultiplex PCR at two research laboratories are shown inTable 1. While the hblDA genes could be determined in134 of the 147 (91.2%) isolates in laboratory 1, thesegenes could be shown to be present in only 128 of the147 isolates (87.1%) in laboratory 2. The nheAB geneswere present in approximately equal numbers of strains,i.e. in 108 vs. 107 of the 147 strains (72.8% vs. 73.5%) inlaboratories 1 and 2, respectively. The cytK-2 gene wasdetected in both departments in 55 of the 147 strains

(37.4%), while none of the strains could be shown tocontain the ces gene.According to the toxin gene profiles of Ehling-Schulz

et al. [16] the majority of strains (79 of 147; 53.7%)encoded the nheAB and hblDA genes and thus belongsto the toxin gene profile C (Table 1) in this study. Inaddition, about 25% strains were found to encode thehblDA and cytK genes which was not defined as a toxingene profile by Ehling-Schulz et al. [16]. The other toxingene profiles F (only the nheAB toxin genes) and G (onlycytk toxin genes) were determined to occur at ca. 9%and ca. 3% in this study (Table 1).The Duopath® test for detection of toxin production

showed a positive result for the Nhe toxin in 122/147 ofthe strains (83.0%), while the Hbl toxin was detected in

Fig. 1 FastTree phylogenetic comparison with whole genome data. A homologous group filtering and a group alignment were performed by PATRIC pipeline[24] and an estimated phylogenetic tree from concatenated alignment sequences was calculated within the type strains B. cereus ATCC 14579T (DSM 31T), B.thuringiensis ATCC 10792T (DSM 2046T), B. toyonensis BCT-7112T (CECT 876T), B. weihenstephanensisWSBC 10204T (DSM 11821T), B. mycoides ATCC 6462T (DSM2048T) and Bacillus pseudomycoides AFS069374T using a FastTree method [25]. Geobacillus thermoglucosidasius DSM 2542T was used as outgroup strain

Fiedler et al. BMC Microbiology (2019) 19:250 Page 3 of 13

Table

1Incide

nces

oftoxinge

neshb

l,nh

e,cytK-2

andcesde

tectionandof

toxins

Nhe

andHbl

expression

Dep

artm

ent/labo

ratory

Presen

ceof

toxinge

nede

tected

inmultip

lex

PCR

Toxinge

neprofile

(classificatio

naccordingto

Ehling-Schu

lzet

al.

[16])

Duo

path®testfortoxin

prod

uctio

n

hblD

nheA

cytK-2

ces

hbl+/nhe+/

cytK+/ces-

(A)

hbl+/nhe+/

cytK−/ces-

(C)

hbl+/nhe−/

cytK+/ces-

-

hbl

−/nhe+/

cytK−/ces-

(F)

hbl−/nhe-

cytK+/ces-

(G)

Nhe

Hbl

Microbiolog

yandBiotechn

olog

y(Kiel)

Labo

ratory

1134/147

(91.2%

)108/147

(73.5%

)55/147

(37.4%

)0/147

(0%)

16/147

(10.9%

)79/147

(53.7%

)39/147

(26.5%

)13/147

(8.8%)

0/147(0%)

122/147

(83.0%

)93/147

(63.3%

)

Safety

andQualityof

FruitandVege

tables

(Karlsruhe

)Labo

ratory

2128/147

(87.1%

)107/147

(72.8%

)55/147

(37.4%

)0/147

(0%)

15/147

(10.2%

)78/147

(53.1%

)35/147

(23.8%

)14/147

(9.5%)

5/147(3.4%)

Com

bine

dresultof

multip

lexPC

RandDuo

path®testfortoxinprod

uctio

n

Microbiolog

yandBiotechn

olog

y(Kiel)

Labo

ratory

155/147

(37.4%

)79/147

(53.7%

)0/147(0%)

13/147

(8.8%)

0/147(0%)

Safety

andQualityof

FruitandVege

tables

(Karlsruhe

)Labo

ratory

255/147

(37.4%

)79/147

(53.7%

)0/147(0%)

13/147

(8.8%)

0/147(0%)

Fiedler et al. BMC Microbiology (2019) 19:250 Page 4 of 13

Fig. 3 Distribution of inhibition zone diameter in the antibiotic disc diffusion test. Each average value was obtained from triplicate test. a erythromycin, btetracycline, c chloramphenicol, d gentamicin. Green indicates susceptible, yellow intermediate and red resistant to antibiotics

Fig. 2 Antibiotic resistance profile of Bacillus strains based on inhibition zone diameter (mm); (red) resistance, (yellow) intermediate, (green) susceptible

Fiedler et al. BMC Microbiology (2019) 19:250 Page 5 of 13

93 (63.3%) of the strains. When the toxin productiondata as determined with the Duopath® test were consid-ered together with the multiplex PCR data, the numberof strains belonging to toxin gene profile A (hbl+, nhe+,cytK+, ces-) increased from approx. 10 to 37.4% ofstrains, while the number of strains belonging to toxinprofile C and F stayed the same at 53% and ca. 9%, re-spectively (Table 1). The Duopath® test was able to in-crease the number of Nhe-positive strains assessed onthe basis on the presence of the nheAB gene, as in themultiplex PCR approx. 73% of strains possessed thenheAB gene, while in the Duopath® test 83% of strainsproduced Nhe. A different situation existed for assess-ment on the incidence of the Hbl toxin, as the multiplexPCR showed 91.2% (laboratory 1) or 87.1% (laboratory2) of strains to possess the hblDA gene, while toxin pro-duction assessment with the Duopath® test showed only63.3% of strains to be positive (Table 1).

Antibiotic resistanceThe B. cereus-group strains were generally resistant tothe β-lactam antibiotics penicillin G (PEN) and cefotax-ime (CTX) (100%), as well as ampicillin (AMP) andamoxicillin/clavulanic acid combination (AMC) (99.3%)(Fig. 2). On the other hand, these strains were generally

susceptible to ciprofloxacin (CIP) (99.3%), chloram-phenicol (CHL) (98.6%), amikacin (AMK) (98.0%), imi-penem (IPM) (93.9%), erythromycin (ERY) (91.8%),gentamicin (GEN) (88.4%) and tetracycline (TET)(76.2%), while 52.4% of strains were susceptible totrimethoprim-sulfamethoxazol (SXT) (Fig. 2).Conversely, only few strains were resistant to the anti-

biotics CHL (1.4%), CIP and IPM (0.7%), ERY (4.1%),GEN (11.6%) and TET (4.1%), while more (36.1%) wereresistant to SXT, as based on the EUCAST breakpoints(Fig. 2). Intermediate susceptible strains were also deter-mined for AMK (2.0%), ERY (4.1%), IMP (5.4%), TET(19.7%) and SXT (11.6%).The inhibition zone ranges (in mm diameter) of the

different antibiotics at their respective concentrationsused in this study were investigated more closely. Suchinhibition zones grouping the isolates together (Fig. 3 a-d) ranged from 17 to 33 mm in the case of ERY, 18 to33mm for TET, 21.5 to 30 mm for CHL and 15 to 31mm for GEN. Isolates exhibiting smaller sized inhibitionzones outside this range, which were also considered tobe resistant according to the EUCAST guidelines [26](for S. aureus), were interesting candidates for investigat-ing the genetic basis of the antibiotic resistance by wholegenome sequencing.

Table 2 Characteristics of whole genome datasets of selective B. cereus s.l. strains

Characteristic Bacillus cereus s.l. isolate

B26 G12 MS12 MS17 MS195 MS464a MS532a MS735

No. of contigs 67 64 368 48 79 41 75 77

Largest contig 524,166 855,854 262,491 733,489 594,878 746,015 1,014,035 657,340

N50 228,834 193,941 44,422 251,250 146,479 249,433 154,418 201,858

GC-content(mol%)

35.27 35.08 34.73 35.31 34.90 35.27 34.70 34.89

Total length

(bp)

5,439,456 5,686,136 6,345,988 5,472,616 6,018,360 5,170,399 6,113,939 5,964,825

MLST (B. cereus) ST-352 unknown ST-15 unknown ST-278 ST-163 unknown ST-72

Species(KmerFinder)

Bacillusweihenstephanensis

Bacilluscereus

Bacillusthuringiensis

Bacillusweihenstephanensis

Bacillustoyonensis

Bacillus cereus Bacillusthuringiensis

Bacillustoyonensis

Antibioticresistance

phenotypeA

AMP, CTX, AMG,PEN, SXT, GM

AMP, CTX,AMG, PEN,SXT

AMP, CTX,AMG, PEN

AMP, CTX, AMG,PEN, SXT

AMP, CTX,AMG, PEN,ERY,

AMP, CTX,AMG, PEN,SXT, GEN

AMP, CTX, AMG,PEN, TET, GEN,ERY, CHL

AMP, CTX,AMG, PEN,ERY,

EnterotoxinPCR

nhe+, hbl+, cytK,ces−

nhe+, hbl+,cytK+, ces−

nhe-, hbl+,cytK+, ces−

nhe+, hbl−, cytK−,ces−

nhe+, hbl+,cytK, ces−

nhe+, hbl−,cytK−, ces−

nhe+, hbl+, cytK+,ces−

nhe+, hbl+,cytK−, ces−

Enterotoxinserology

Nhe+, Hbl− Nhe+, Hbl+ Nhe+, Hbl+ Nhe+, Hbl− Nhe+, Hbl+ Nhe+, Hbl− Nhe+, Hbl+ Nhe+, Hbl+

Enterotoxingenes identifiedon genome

nhe and hbl genes nhe, hbland cytKgenes

nhe, hbl andcytK genes

nhe and hbl genes nhe and hblgenes

nhe and hblgenes

nhe, hbl and cytKgenes

nhe and hblgenes

Species identification was carried out with KmerFinder (PATRIC) which examined the number of co-occurring k-mers (substrings of k nucleotides in genomesequence data) and MLST was done by PATRIC to assign the strains to clonal lineages. AbbreviationsA: amikacin (AMK), ampicillin (AMP), amoxicillin/clavulanicacid (AMG), cefotaxime (CTX), chloramphenicol (CHL), ciprofloxacin (CIP), erythromycin (ERY), gentamicin (GEN), imipenem (IPM), penicillin G (PEN), tetracycline(TET) and trimethoprim-sulfamethoxazole (SXT)

Fiedler et al. BMC Microbiology (2019) 19:250 Page 6 of 13

Whole genome sequencingThe genomes of selected antibiotic-resistant strains weresequenced (Table 2) and genome sizes ranged from 5.17to 6.35 Mbp, while the mol% G + C content ranged from34.70 to 35.31. Liu et al. [14] reported that within a col-lection of 224 B. cereus s.l. strains, the genome sizesranged from 4.09 to 7.09 Mbp with an average 5.77Mbp, which agreed well with the genome sizes deter-mined in our study. Furthermore, the mol% G + C valuesof these genomes were reported to range from 34.5 to36.7% [14]. Again, the data from the genome sequencesobtained for the strains in this study agree well, as theyfit within this range. Sequencing the genomes of the se-lected isolates confirmed the results of the multiplexPCR (in combination with serological testing) for B. ce-reus s.l. enterotoxin genes, as the genes detected on thegenomes with automated annotation were the same en-terotoxin profile as detected by combined multiplex PCRand enterotoxin serological testing results (Table 2).Concerning the results of the enterotoxin gene ana-

lyses by multiplex PCR and serological testing, some in-teresting discrepancies occurred in the results. For

example, in strain MS12 no PCR product could be ob-tained in the nheAB PCR, yet a signal was obtainedusing the Duopath® test and genome sequencing showedthe genes to be present. Clearly, this must havedepended on polymorphisms in the nheAB gene nucleo-tide sequence, that resulted in no or insufficient primerbinding. Indeed, the nheAB forward primer (NA2-F)showed a T/C mismatch for strain MS12, which waspresent in the PCR primer target sequence of only thisstrain (Additional file 2: Figure S2a). Furthermore, forstrains B26 and MS17, the Hbl toxin results were some-what confusing. Strain B26 showed a PCR signal in themultiplex PCR for the hblDA genes, while strain MS17did not. The genome sequence of both strains B26 andMS17 showed the Hbl operon to be present. Thus, thelack of hblDA signal in PCR for strain MS17 can be ex-plained by a G/A primer mismatch for primer hbl F(HD2 F) (Additional file 2: Figure S2b). Moreover, in theserological assay, the Hbl toxin (HblC subunit) couldnot be detected for both strains B26 and MS17, indicat-ing that possibly the amino acid sequence differences inthis gene hblC in these two strains did not allow the

Fig. 4 The amino acid sequences of the hblC gene of Bacillus cereus s.l. strains were clustered in this study. Most of strains reacted Hbl-positivewith the Duopath test but two strains (B26 and MS17) were Hbl-negative. The phylogenetic analysis was carried out using the Jukes-Cantormodel for genetic distance and the unweighted pair-group method using arithmetic averages (UPGMA) algorithm of the Geneious® program.The branch length and branch support values indicate the units of substitutions per site of the sequence alignment

Fiedler et al. BMC Microbiology (2019) 19:250 Page 7 of 13

antibody to react with the toxin. Performing a phylogen-etic tree with the amino acid sequence of the HblC pro-teins showed a clear separation of B26 and MS17 fromthe other six strains (Fig. 4). Alternatively, it is knownthat the detection limit of the Duopath® test kit is 20 ng/ml and possibly the strain did not produce sufficienttoxin to show a serological reaction.The genome sequencing showed that all sequenced

strains contained a gene encoding a fosfomycin resistanceprotein FosB, a broad specificity multidrug efflux pumpYkkC and a gene encoding a putative streptogramin O-acetyltransferase involved in streptogramin A resistance(Additional file 3: Table S1). Genes encoding β-lactam anti-biotic resistances present on the genome generally includedgenes for class A β-lactamases (EC 3.5.2.6) and class B (sub-class B1) β-lactamases (Additional file 3: Table S1).Of the eight sequenced strains, three were resistant to

erythromycin. These three strains possessed each two ormore genes associated with macrolide resistance (macro-lide 2′-phosphotransferase of the Mph(B) family, puta-tive macrolide 2′-phosphotransferase and ABC-F typeribosomal protection protein Lsa(B)). However, otherstrains were susceptible to erythromycin, even thoughthe same genes were present (Additional file 3: TableS1). Strain MS532a possessed a tet(45) tetracycline re-sistance gene, encoding an efflux pump (Additional file3: Table S1). This strain showed a tetracycline resistantphenotype and displayed an inhibition zone of only 8.5mm in the disc diffusion test. Six of the eight sequencedstrains contained a chloramphenicol O-acetyltransferasegene (Additional file 3: Table S1), but of these only one(MS532a) displayed a chloramphenicol-resistant pheno-type. All strains contained an aminoglycoside 6-nucleotidyltransferase gene, as well as a putative genewith weak similarity to an aminoglycoside N(3)-acetyl-transferase. However, none of the strains showed resist-ance to amikacin and only three strains displayedgentamicin resistance.

DiscussionThe 16S rRNA gene sequence showed that bacteria inthis study isolated from B. cereus selective agar allbelonged to the B. cereus s.l. group. The intention of thisstudy, however, was not to accurately distinguish be-tween the different members of the B. cereus s.l. group,but only to confirm that the isolates were members ofthis group. To address the importance of these bacteriain the context of food safety, it was more important todetermine whether they possessed gene(s) for toxin pro-duction. Nevertheless, our phylogenomic assessment ofthe eight sequenced strains showed that strains fromvegetables in this study consisted of a mix of species, in-cluding B. toyonensis, B. cereus/thuringiensis and B.mycoides/weihenstephanensis (Fig. 1).

Our multiplex PCR results showed that a large propor-tion (91.2%) of the B. cereus-group strains from freshproduce possessed the hbl gene, while 73.5% of thestrains possessed the nhe and 37.4% possessed the cytKgenes. This was slightly lower than reported for Bacilluscereus s.l. group strains isolated from dried spices andherbs in Germany [5]. In that investigation 94.9% ofstrains possessed the hbl gene, while the nhe gene andthe cytK genes were present in 96.6 and 50.8% of strains,respectively [5]. Other reports have also shown that thenhe gene sequence is detected at high frequency [27, 28].Park et al. [29] detected the nhe, hbl and cytK genes in92, 93 and 55% of B. cereus strains from cereals, respect-ively. The lower prevalence of cytK genes observed inthis study was also in accordance with other studies [5,29, 30]. Bacillus cereus strains possessing the cytK genehave a high pathogenic potential, as the cytK gene wasfrequently detected in strains associated with diarrheaand food poisoning [29, 31]. The results for multiplexPCR from laboratory 1 and laboratory 2 differed onlyslightly, which may be explained by the small variationsin PCR protocol, PCR chemicals (including type of poly-merase used) as well as PCR equipment.Our combined multiplex PCR and Duopath® tests sug-

gested that half of the strains (53.7%) were positive forboth Hbl and Nhe, while more than a third (37.4%) werepositive for all three toxins Nhe, Hbl and CytK. In thestudy of Frentzel et al. [5] 49.2% of strains possessed allthe three hbl, nhe and cytK genes, while 45.8% possessedthe genes for hbl and nhe, only. In the study of Frentzelet al. [5] a single strain was found to possess the ces genefor cereulide production, while this gene could not befound among any of the strains in this study.These results suggested that toxinogenic B. cereus s.l.

group organisms occur in spices, herbs and fresh pro-duce in Germany and that these commonly containgenes for diarrheal toxin production, while the occur-rence of strains producing emetic toxin is rather low.Our study confirmed previous reports that the nhe andthe hbl toxin genes were most prevalent, and the cytKenterotoxin genes occurred at a lower incidence thanthese two [5, 29]. Böhm et al. [32] showed a strictly ver-tical inheritance for the nhe gene, which may explainwhy this gene is more prevalently occurring in B. cereuss.l. strains, while there was ample evidence for duplica-tion of hbl, horizontal transfer of hbl and cytK, as well asfrequent deletion of both toxins. The latter may explainthe lower occurrence of the cytK toxin.In our previous investigation on the microbiological

quality and safety of fresh produce from the Germanmarket, we reported that B. cereus s.l. counts in freshproduce such as cucumbers, carrots, herbs, leaf lettuceand mixed salad leaves obtained directly from retailclose to the end of the use by date are quite low, with

Fiedler et al. BMC Microbiology (2019) 19:250 Page 8 of 13

mean counts ranging from 1.26 × 101 to 5.1 × 101 cfu/g[2]. Becker et al. [6] and Frentzel et al. [5] reported B. ce-reus s.l. counts in salads and herbs, and in dried spicesand herbs from the German market to range from ca.102 to 104 cfu/g and < 8.0 × 101 to 1.6 × 103 cfu/g, re-spectively. These findings are furthermore in agreementwith reports of Banerjee and Sarkar [33], Kneifel andBerger [34] and Sagoo et al. [35], who reported presenceof only low (less than 1 × 104 cfu/g) levels of B. cereus-group strains in condiment samples. Bacillus cereus diar-rheal toxins are formed in the human gut after con-sumption of cells or spores in quantities of usually morethan 1 × 105 cfu/g [1, 5, 36]. These data, therefore, indi-cate that B. cereus-group strains containing mostlyenterotoxin genes are present in dried herbs andspices, as well as fresh produce on the German retailmarket and that they do pose an inherent risk toconsumers. However, usually these occur at too lownumbers to cause intoxication.The increased occurrence of multiple antibiotic resist-

ance is currently a concern for food processors and publichealth officials. As fresh produce can be eaten without fur-ther heat treatment, there is a concern that antibiotic-resistant bacteria may survive gastrointestinal passage andmay complicate the treatment in the very young or oldconsumers or persons with lowered immune functionafter infection. For this reason, we also investigated the in-cidence of antibiotic resistances in the B. cereus s.l. isolatesobtained from fresh produce. Bacillus cereus group strainsare known to be typically resistant towards β-lactam anti-biotics as a result of production of β-lactamase enzymes[29, 37–39]. This was confirmed in this study, whichshowed that strains were generally resistant to penicillinG, ampicillin, cefotaxime and the combination of amoxi-cillin with the clavulanic acid as β-lactamase inhibitor.Similar to the results of other studies [29, 38, 39], thesebacteria were generally susceptible towards other antibi-otics, including tetracycline, erythromycin, chlorampheni-col, gentamicin, ciprofloxacin and imipenem. A fewstrains did, however, show reduced susceptibility to genta-micin, tetracycline and erythromycin (Figs. 2 and 3). Resis-tances to tetracycline and erythromycin in these bacteriahave previously been noted to occur in the United Statesand Europe [38].The genomes of three erythromycin-resistant and five

erythromycin-susceptible strains were sequenced. Bothresistant and susceptible strains contained a putativemacrolide 2′-phosphotransferase gene, but it was notedthat all resistant strains contained a gene encoding anABC-F type ribosomal protection protein Lsa(B) (Add-itional file 3: Table S1) whose protein protects the ribo-some from the antibiotics action. The three resistantstrains showed markedly reduced inhibition zone diame-ters of 10 mm (MS 532a), 14.25 mm (MS 195) and 14.5

mm (MS 735) (Fig. 3a), which were noticeably smallerthan the EUCAST suggested breakpoint diameter for S.aureus used in this study of < 18 mm. Therefore, thisvalue for erythromycin could be decreased to < 15mmas a new breakpoint for Bacillus cereus s.l. according tothe results of this study (Fig. 3a).The genome of the tetracycline-resistant strain MS

532a, which showed an 8.5 mm diameter zone of inhib-ition, was sequenced and a gene for a tetracycline resist-ance efflux pump tet(45) was identified that couldexplain the resistance mechanism (Additional file 3:Table S1). As there are no EUCAST criteria for deter-mining susceptibility of Bacillus strains, the EUCASTcriteria for S. aureus were used in this study. A break-point of < 19mm for tetracycline is specified in this caseto indicate resistance, which in our opinion is not suit-able for Bacillus, as the breakpoints of non-resistantstrains ranged from 18 to 33 mm (Fig. 3a). Thus, wewould suggest that for Bacillus cereus s.l. group strainsthis value should be < 18mm or even less.Strains were also chosen for genome sequencing that

appeared to be chloramphenicol or gentamicin resistant.Although the small diameters of the inhibition zonesindicated the strains to be clearly resistant, genomesequencing did not detect a specific genetic determinantthat could explain the resistance fully. In the case ofchloramphenicol, all sequenced strains possessed achloramphenicol O-acetyltransferase of the CatA15/A16family (Additional file 3: Table S1), yet only one strain(MS532a) was phenotypically resistant (Figs. 2 and 3c).Should this gene be responsible for conferring resistance,the gene must have been present in an inactivatedpseudogene form in the sensitive strains. In the case ofgentamicin resistance, the strains B26, MS532a andMS464a were phenotypically resistant (Figs. 2 and 3d),yet putative aminoglycoside 6-nucleotidyltransferasegenes were detected in all sequenced genomes, of which3 strains did not show a resistance. Again, this may bepossibly a result of the presence of pseudogenes withmutations in susceptible strains or the resistances mayrely on other unknown genes and mechanisms.

ConclusionThe results of our previous study as well as this studyshow that B. cereus s.l. strains can occur in fresh vegeta-bles on the German retail markets at levels of up to ca.log 3.0log 4.0 cfu/g [2, 6] and that these contain differentcombinations of enterotoxin genes that are known to beresponsible for causing diarrhea. The methods for detec-tion of these toxins, i.e. multiplex PCR, genome sequen-cing and serological toxin detection were shown to leadto somewhat differing results, probably as a result oftoxin gene sequence polymorphisms. Thus, a combin-ation of methods should be applied to obtain an accurate

Fiedler et al. BMC Microbiology (2019) 19:250 Page 9 of 13

assessment of the presence of toxin genes in B. cereuss.l. strains. None of the strains were detected to containthe gene for the cereulide toxin. Generally, the levels offresh vegetable contamination with B. cereus were pre-sumably too low to cause toxin production and food-borne disease. This probably explains why thereported incidences of foodborne disease caused bythese bacteria after consumption of fresh vegetablesin Germany are rather low. However, these bacteriacan be consistently isolated from such products. Inthe case of ready-to-eat products, this shows thatthere is an inherent risk associated with these prod-ucts if storage temperatures are abused or if theproducts become contaminated with higher B. cereuss.l. loads. Moreover, the effect of these toxin-producers on vulnerable consumers is not yet known.For these reasons, a close vigil should be kept on theoccurrence and levels of B. cereus s.l. in such freshvegetables, especially ready-to-eat products. Bacilluscereus s.l. are also generally sensitive towards antibi-otics. It could be determined in this study that anti-biotic resistance breakpoint values supplied for S.aureus strains in EUCAST were suitable for many,but not all antibiotics tested against B. cereus groupstrains. The data generated in this study could be thebasis for further studies of such resistance break-points. Nevertheless, our results indicate that resis-tances towards antibiotics such as tetracycline,erythromycin and aminoglycosides, can occur. Onestrain in this study was multiple resistant towards β-lactam antibiotics, aminoglycoside, tetracycline anderythromycin. The presence of these antibiotic resist-ance genes in these gram-positive, Bacillus microor-ganisms therefore appears to mirror the currentoccurrence of such resistance genes in the vegetable

environment. Furthermore, it thus represents a pos-sibly less important source of potentially transferableresistance genes in the food chain.

MethodsBacterial strainsA total of 147 B. cereus s.l. strains were isolated from 137fresh vegetable samples, i.e. cucumbers (n = 6), carrots (n =11), herbs (n = 22), salad leaves (either lollo bionda, lamb’slettuce, spinach or arugula) (n = 19) and ready-to-eat mixedsalads (n = 89), purchased from retail stores, local marketsand farmyard sales in Germany. The isolates and the prod-ucts from which they were isolated are shown in Table 3.B. cereus s.l. strains (n = 73) were selectively isolated on Ba-cillus-cereus-Agar (PEMBA; Sifin, Berlin, Germany) at 30 °Cfor 48 h [2], while 74 further strains were isolated on Bacil-lus cereus Rapid Agar (BACARA; bioMérieux, Nürtingen,Germany) at 30 °C for 48 h (n = 74 isolates, see Table 3).Some of the 73 strains isolated on PEMBA agar were previ-ously identified and described [2].

Phenotypic characterization of Bacillus cereus-groupisolatesAll 147 isolates were tested for growth at different tem-peratures (4, 7, 10 and 50 °C) for 10 days in Caso orStandard 1 Broth (Merck, Darmstadt, Germany). Ahemolysis test was performed by streaking onto com-mercially prepared Columbia agar (Oxoid, Wesel,Germany) plates containing 7% sheep’s blood and incu-bating for 24 h at 30 °C. Plates were examined for α-, β-and ɣ-hemolysis.

Antibiotic susceptibility testingAntimicrobial susceptibility of the B. cereus s.l. specieswas tested using the disc diffusion method based on the

Table 3 Origin of Bacillus cereus-groupstrains investigated in this study

Source of isolation and no. of samples (137) Isolatesa Total no. ofisolates (147)

Carrots (11) M5, M6, M10, M14, M15, M18, M19, M20, M26, M27, M39 11

Cucumbers (6) G2, G8, G12, G13, G32, G36 6

Herbs (22) K1, K2, K10, K11, K13, K14, K16, K17, K20, K21, K22, K23, K28, K30, K31, K33, K34, K35,K37, K38, K39, K40

22

Salad leaves (either one of the following: lollobionda, lamb’s lettuce, spinach, arugula) (19)

B2, B8, B15, B16, B19, B20, B24, B25, B26, B27, B29, B30, B31, B32, B34, B36, B37, B38,B40

19

Mixed salad leaves (79) MS3, MS4, MS6, MS9, MS12, MS14, MS17, MS18, MS19, MS25, MS27, MS29, MS33,MS36, MS38, MS79, MS135a, MS135b, MS136a, MS136b, MS138, MS139, MS142,MS143, MS147, MS190, MS191, MS192, MS194, MS195, MS360a, MS360b, MS361,MS362, MS456, MS457, MS460, MS461, MS463a, MS463b, MS464a, MS464b, MS467,MS468a, MS468b, MS469, MS470, MS471, MS525, MS526a, MS527, MS529, MS530,MS531, MS532a, MS532b, MS567, MS568, MS569, MS575, MS578a, MS578b, MS579,MS581, MS582, MS586, MS588, MS689, MS690, MS691a, MS691b, MS693, MS728a,MS728b, MS729, MS730, MS733, MS734, MS735, MS756, MS758, MS788, MS789,MS790, MS791, MS792, MS793, MS794, MS795

89

a isolates marked in italic type were isolated on PEMBA agar and were previously reported regarding their 16S rRNA gene based identification in the study ofFiedler et al. [2]. All other strains were isolated on BACARA agar

Fiedler et al. BMC Microbiology (2019) 19:250 Page 10 of 13

European Committee on Antimicrobial SusceptibilityTesting (EUCAST) guidelines [26]. A total of 12 differ-ent antibiotic discs (Oxoid) containing the antibioticsamikacin (AMK 30 μg), ampicillin (AMP 10 μg), amoxi-cillin/clavulanic acid (AMC 20/10 μg), cefotaxime(CTX 30 μg), chloramphenicol (CHL 30 μg), ciproflox-acin (CIP 5 μg), erythromycin (ERY 15 μg), gentamicin(GEN 10 μg), imipenem (IPM 10 μg), penicillin G(PEN 10 μg), tetracycline (TET 30 μg) andtrimethoprim-sulfamethoxazole (SXT 25 μg) were usedfor susceptibility testing on Mueller Hinton Agar(Merck). Susceptibility of the isolates to the antibi-otics was determined by measuring the zone of inhib-ition, and these were interpreted as sensitive,intermediate or resistant according to EUCAST [26]criteria for Staphylococcus (S.) aureus (except imipe-nem and amoxicillin/clavulanic acid), as no Bacillus-specific criteria for disc diffusion antibiotic-resistanceassays have been defined either by EUCAST or Clin-ical & Laboratory Standards Institute guidelines(CSLI). For imipenem, the Enterococcus spp.(EUCAST) and for amoxicillin/clavulanic acid, theEnterobacteriales (EUCAST) breakpoints wereadopted. All antibiotic resistance determinations werecarried out in duplicate on different days, using dif-ferent agar batches and different overnight-subcultures. For ambiguous results (difference be-tween duplicates > 4 mm) a third repetition was per-formed. Data were graphically presented using JMP(v14, SAS software, Cary, USA).

Duopath® test for detection of Nhe and Hbl bacterialtoxin formationThe Duopath® Cereus Enterotoxins test (Merck) test detectsthe NheB component of the non-haemolytic three-component enterotoxin (Nhe) and the HblC component ofthe three-component haemolysin BL (Hbl) [40] and was per-formed using overnight Bacillus isolates broth cultures. TheBacillus isolates were streaked onto mannitol egg yolk poly-myxin (MYP) agar (Oxoid). After incubation at 30 °C for 18to 24 h, one to three colonies were picked and suspended in1ml of casein hydrolysate-glucose-yeast extract broth (CGY)(Merck). After 4 h incubation at 37 °C, a 150 μl aliquot of theenrichment was used to inoculate the Duopath® Cereus En-terotoxins lateral flow immunoassay and incubated for 30min at room temperature. The test was considered positive ifa red line became visible within 30min in both the test (Nheand /or Hbl) and control zones and negative if a line ap-peared only in the control zone.

DNA extraction and 16S rRNA gene sequencingThe cultures were grown overnight in Caso Broth orStandard 1 Broth, and DNA was extracted either usingthe ZR Fungal/Bacterial DNA MiniPrep™ (Zymo

Research, Freiburg, Germany) according to the manufac-turer’s instructions or alternatively according to Pitcheret al. (1989). DNA concentration was quantified usingNanoDrop 2000 (Thermo Scientific, Waltham, USA) orusing a Qubit fluorometer (Invitrogen, Darmstadt,Germany).For genotypic characterization, the 86% complete 16S

rRNA gene sequence of the isolates (ca. 1342 bp) wasobtained as previously described by Fiedler et al. [2] andDanylec et al. [41]. The sequences were compared andphylogenetic analyses were performed by fast algorithmand unweighted pair group method with arithmeticmean (UPGMA) clustering using BioNumerics (v7.6,Applied Maths, Saint-Martens-Latem, Belgium).

PCR screening for enterotoxin genesBacillus cereus s.l. strains toxin genes were detected bymultiplex PCR according to Ehling-Schulz et al. [16]. Thismultiplex PCR was designed to simultaneously amplifythe enterotoxin genes hblDA, nheAB, cytK-2 and ces withPCR product sizes of 1091 bp, 766 bp, 421 bp and 1271 bp,respectively. Suitable reference strains from the DSMZculture collection were used as positive controls. Themultiplex PCR reactions were done independently at twodifferent laboratories at different departments of the MaxRubner-Institut, i.e. the Department of Safety and Qualityof Fruit and Vegetables located in Karlsruhe and the De-partment of Microbiology and Biotechnology located inKiel, using the method of Ehling-Schulz et al. [16] withslight variations (see Additional file 4: Table S2). The PCRproducts were separated by gel electrophoresis, stainedwith ethidium bromide (Merck) and investigated using aUV transilluminator (VWR).

Whole genome sequencing and analysesThe sequencing library was prepared with an IlluminaNextera XT library prep kit (Illumina, San Diego,USA) and sequenced on a MiSeq (Illumina) sequencerwith 2 × 250 paired-end reads. The reads were denovo assembled using SPAdes version 3.11.1 [42].The genome sequence was annotated and assigned toB. cereus multi locus sequence types using thePATRIC database [43]. The acquired antibiotic resist-ance genes were identified using the ResFinder server(v. 3.0) [44]. Genome and protein sequences were an-alyzed and comparisons were visualized using Gen-eious (v9.0.5, Biomatters Limited, New Zealand). Inorder to construct a phylogenetic tree based on gen-ome sequences, the annotated amino acid sequenceencoded by each genome were extracted and utilizedby the PATRIC server. A homologous group filteringand a group alignment were performed by this pipe-line and an estimated phylogenetic tree from

Fiedler et al. BMC Microbiology (2019) 19:250 Page 11 of 13

concatenated alignment sequences was calculated withthe outgroup type strain Geobacillus thermoglucosida-sius DSM 2542T using a FastTree method [25].

Supplementary informationSupplementary information accompanies this paper at https://doi.org/10.1186/s12866-019-1632-2 .

Additional file 1: Figure S1. Cluster analysis of 16S rRNA genesequences of 147 Bacillus cereus s.l. and selected type strains (B. anthracisATCC 14578T, B. cereus DSM 31T, B. cytotoxicus DSM 22905T, B. mycoidesDSM 2048T, B. pseudomycoides DSM 12442T, B. subtilis DSM 10T, B.thuringiensis DSM 2046T, B. toyonensis CECT 876T, B. weihenstephanensisDSM 11821T). Strains isolated in Kiel and Karlsruhe are labelled green andred, respectively. Reference strains are labelled in blue. Fast algorithm assimilarity coefficient and UPGMA were used. Due to the high number ofstrains the dendrogram was condensed.

Additional file 2: Figure S2. (a) Primer mismatches nheA: Analysis ofthe nheA region targeted by the PCR Primer NA2F. The yellow boxindicates a nucleotide mismatch and the nucleotide positions areindicated above the primer sequence. Inosine was used as a degeneratebase and was marked with an I in the primer sequence. (b) Primermismatches hblD: Analysis of the hblD region targeted by PCR withPrimer HD2F. A yellow box indicates a nucleotide mismatch and thenucleotide positions are indicated above the primer sequence. Inosinewas used as a degenerate base and was marked with an I in the primersequence.

Additional file 3: Table S1. Antibiotic resistance genes identified onthe genomes of B. cereus-group strains using the PATRIC [43] database(bold highlights acquired resistance genes identified by ResFinder [45]).

Additional file 4: Table S2. Multiplex PCR conditions used for toxingene amplification.

AbbreviationsATCC: American Type Culture Collection; B.: Bacillus; bp: Base pair;ces: cereulide; CLSI: Clinical and Laboratory Standards Institute guidelines;cytK: cytotoxin K; DNA: deoxyribonucleic acid; DSMZ: Deutsche Sammlungvon Mikroorganismen und Zellkulturen; EUCAST: European Committee onAntimicrobial Susceptibility Testing; hbl: hemolysin BL; mbp: mega base pair;MLST: multilocus sequence typing; mm: millimeter; nhe: non-hemolyticenterotoxin; PCR: polymerase chain reaction; PEMBA: Polymyxin pyruvateegg yolk mannitol bromothymol blue agar; rRNA: ribosomal ribonucleic acid;S.: Staphylococcus; s.l.: sensu lato; s.s.: sensu stricto; UPGMA: unweighted pairgroup method (with) arithmetic mean

AcknowledgementsTechnical assistance of Gesa Gehrke, Luisa Martinez, Jessica Obermeyer,Adrian Prager, Lilia Rudolf, Karin Schavitz and Lena Schmid is kindlyacknowledged.

Authors’ contributionsGF, BB, MH and CF designed experiments. GF, BB, JK, DS and MH supervisedthe laboratory testing. CS, EOI, EB, DS, GSC and GF carried outmicrobiological analyses, biochemical and molecular genetic studies. GF, CS,BB, DS, GSC, MH and CF contributed to the work with data analysis andinterpretation of results. CF and GF drafted the manuscript, while all otherauthors reviewed and edited. All authors read and approved the finalmanuscript.

FundingThis study was supported by the German Federal Ministry of Food andAgriculture and partially by the Alexander von Humboldt Foundation withregards to antibiotic resistance testing by EOI. The work was conducted withinthe project “Humanpathogene in der pflanzlichen Erzeugung: Status quo,Dekontamination, Eintragswege und Einfluss der Lagerungsbedingungen.”. EOIgratefully acknowledges the Alexander von Humboldt Foundation’s GeorgForster Stipendium in the framework of the Research Group LinkageProgramme, funded by the Federal Foreign Office, the Federal Ministry of

Education and Research and the Federal Ministry for Economic Cooperationand Development, for Experienced Researchers for financing his research stayat the Max Rubner-Institut (MRI).The funding bodies had no role in study design, data collection and analysis,interpretation of data, decision to publish, or preparation of the manuscript.

Availability of data and materialsThe data-sets analyzed during the current study are available from the corre-sponding author on reasonable request. The Bacillus cereus s.l. genome se-quences (B26, G12, MS12, MS17, MS195, MS464a, MS532a and MS735) areavailable on the NCBI Genome Database under the accession numbersSJQA00000000, SJQB00000000, SJQC00000000, SJQD00000000,SJQE00000000, SJQF00000000, SJQG00000000 and SJQH00000000,respectively.

Ethics approval and consent to participateNot applicable. Samples and bacterial isolates used in the present study donot require ethical approval or owner consent.

Consent for publicationNot applicable. No personal data were collected in the context of this study.

Competing interestsThe authors declare that they have no competing interests.

Author details1Department of Microbiology and Biotechnology,Hermann-Weigmann-Straße 1, 24103 Kiel, Germany. 2Department of Safetyand Quality of Fruit and Vegetables, Max Rubner-Institut, Federal ResearchInstitute of Nutrition and Food, Haid-und-Neu-Straße 9, 76131 Karlsruhe,Germany. 3Present Address: Department of Microbiology, Faculty of LifeSciences, University of Benin, Private Mail Bag 1154, Benin City 30001, Nigeria.

Received: 28 February 2019 Accepted: 29 October 2019

References1. EFSA. The European Union summary report on trends and sources of

zoonoses, zoonotic agents and food-borne outbreaks in 2016. EFSA J. 2017;15(12):e05077.

2. Fiedler G, Kabisch J, Bohnlein C, Huch M, Becker B, Cho GS, Franz C.Presence of human pathogens in produce from retail Markets in NorthernGermany. Foodborne Pathog Dis. 2017;14(9):502–9.

3. Kotzekidou P. Microbiological examination of ready-to-eat foods and ready-to-bake frozen pastries from university canteens. Food Microbiol. 2013;34(2):337–43.

4. Kim YJ, Kim HS, Kim KY, Chon JW, Kim DH, Seo KH. High occurrence rateand contamination level of Bacillus cereus in organic vegetables on Sale inretail markets. Foodborne Pathog Dis. 2016;13(12):656–60.

5. Frentzel H, Thanh MD, Krause G, Appel B, Mader A. Quantification anddifferentiation of Bacillus cereus group species in spices and herbs by real-time PCR. Food Control. 2018;83:99–108.

6. Becker B, Stoll D, Schulz P, Kulling S, Huch M. Microbial contamination oforganically and conventionally produced fresh vegetable salads and herbsfrom retail Markets in Southwest Germany. Foodborne Pathog Dis. 2018;16(4):269–75.

7. EFSA. Risks for public health related to the presence of Bacillus cereus andother Bacillus spp. including Bacillus thuringiensis in foodstuffs. EFSA J. 2016;14(7):4524.

8. Posfay-Barbe KM, Schrenzel J, Frey J, Studer R, Korff C, Belli DC, Parvex P,Rimensberger PC, Schappi MG. Food poisoning as a cause of acute liverfailure. Pediatr Infect Dis J. 2008;27(9):846–7.

9. Shiota M, Saitou K, Mizumoto H, Matsusaka M, Agata N, Nakayama M, KageM, Tatsumi S, Okamoto A, Yamaguchi S, et al. Rapid detoxification ofcereulide in Bacillus cereus food poisoning. Pediatrics. 2010;125(4):e951–5.

10. Naranjo M, Denayer S, Botteldoorn N, Delbrassinne L, Veys J, Waegenaere J,Sirtaine N, Driesen RB, Sipido KR, Mahillon J, et al. Sudden death of a youngadult associated with Bacillus cereus food poisoning. J Clin Microbiol. 2011;49(12):4379–81.

11. Guinebretiere MH, Velge P, Couvert O, Carlin F, Debuyser ML, Nguyen-TheC. Ability of Bacillus cereus group strains to cause food poisoning varies

Fiedler et al. BMC Microbiology (2019) 19:250 Page 12 of 13

according to phylogenetic affiliation (groups I to VII) rather than speciesaffiliation. J Clin Microbiol. 2010;48(9):3388–91.

12. Guinebretiere MH, Auger S, Galleron N, Contzen M, De Sarrau B, De BuyserML, Lamberet G, Fagerlund A, Granum PE, Lereclus D, et al. Bacilluscytotoxicus sp. nov. is a novel thermotolerant species of the Bacillus cereusgroup occasionally associated with food poisoning. Int J Syst Evol Microbiol.2013;63(Pt 1):31–40.

13. Jimenez G, Urdiain M, Cifuentes A, Lopez-Lopez A, Blanch AR, Tamames J,Kampfer P, Kolsto AB, Ramon D, Martinez JF, et al. Description of Bacillustoyonensis sp. nov., a novel species of the Bacillus cereus group, andpairwise genome comparisons of the species of the group by means of ANIcalculations. Syst Appl Microbiol. 2013;36(6):383–91.

14. Liu Y, Lai Q, Goker M, Meier-Kolthoff JP, Wang M, Sun Y, Wang L, Shao Z.Genomic insights into the taxonomic status of the Bacillus cereus group. SciRep. 2015;5:14082.

15. Schoeni JL, Wong AC. Bacillus cereus food poisoning and its toxins. J FoodProt. 2005;68(3):636–48.

16. Ehling-Schulz M, Guinebretiere MH, Monthan A, Berge O, Fricker M, Svensson B.Toxin gene profiling of enterotoxic and emetic Bacillus cereus. FEMS Microbiol Lett.2006;260(2):232–40.

17. Batchoun R, Al-Sha'er AI, Khabour OF. Molecular characterization of Bacilluscereus toxigenic strains isolated from different food matrices in Jordan.Foodborne Pathog Dis. 2011;8(11):1153–8.

18. Kim JB, Kim JM, Cho SH, Oh HS, Choi NJ, Oh DH. Toxin genes profiles andtoxin production ability of Bacillus cereus isolated from clinical and foodsamples. J Food Sci. 2011;76(1):T25–9.

19. FAO FaAOotUN. The FAO action plan on antimicrobial resistance: 2016–2020. Rome: Food and Agriculture; 2016.

20. WHO: Global strategy on people-centred and integrated health services.Geneva: World Health Organization; 2015 (WHO/HIS/SDS/2015.6; http://www.who.int/servicedeliverysafety/areas/people-centred-care/global-strategy/en/. accessed 9 June 2015). 2015.

21. O'Neill J. Nat Rev Drug Discov. 2016;15:526.22. Citron DM, Appleman MD. In vitro activities of daptomycin, ciprofloxacin,

and other antimicrobial agents against the cells and spores of clinicalisolates of Bacillus species. J Clin Microbiol. 2006;44(10):3814–8.

23. Jensen LB, Baloda S, Boye M, Aarestrup FM. Antimicrobial resistance amongPseudomonas spp. and the Bacillus cereus group isolated from Danishagricultural soil. Environ Int. 2001;26(7–8):581–7.

24. Wattam AR, Gabbard JL, Shukla M, Sobral BW. Comparative genomicanalysis at the PATRIC, a bioinformatic resource center. Methods Mol Biol.2014;1197:287–308.

25. Price MN, Dehal PS, Arkin AP. FastTree 2--approximately maximum-likelihood treesfor large alignments. PLoS One. 2010;5(3):e9490.

26. EUCAST: The European Committee on Antimicrobial Susceptibility Testing.Breakpoint tables for interpretation of MICs and zone diameters. Version 6.0,2016. http://www.eucast.org. 2016. Accessed 19 Apr 2018.

27. Rusul G, Yaacob NH. Prevalence of Bacillus-Cereus in selected foods anddetection of enterotoxin using Tecra-via and Bcet-Rpla. Int J Food Microbiol.1995;25(2):131–9.

28. Anderson Borge GI, Skeie M, Sorhaug T, Langsrud T, Granum PE. Growthand toxin profiles of Bacillus cereus isolated from different food sources. IntJ Food Microbiol. 2001;69(3):237–46.

29. Park YB, Kim JB, Shin SW, Kim JC, Cho SH, Lee BK, Ahn J, Kim JM, Oh DH.Prevalence, genetic diversity, and antibiotic susceptibility of Bacillus cereusstrains isolated from rice and cereals collected in Korea. J Food Prot. 2009;72(3):612–7.

30. McKillip JL. Prevalence and expression of enterotoxins in Bacillus cereus and otherBacillus spp., a literature review. Antonie Van Leeuwenhoek. 2000;77(4):393–9.

31. Guinebretiere MH, Broussolle V, Nguyen-The C. Enterotoxigenic profiles offood-poisoning and food-borne Bacillus cereus strains. J Clin Microbiol.2002;40(8):3053–6.

32. Bohm ME, Huptas C, Krey VM, Scherer S. Massive horizontal gene transfer,strictly vertical inheritance and ancient duplications differentially shape theevolution of Bacillus cereus enterotoxin operons hbl, cytK and nhe. BMCEvol Biol. 2015;15:246.

33. Banerjee M, Sarkar PK. Microbiological quality of some retail spices in India.Food Res Int. 2003;36(5):469–74.

34. Kneifel WEB. Microbiological criteria of random samples of spices and herbsretailed on the Austrian market. J Food Prot. 1994;57(10):893–901.

35. Sagoo SK, Little CL, Greenwood M, Mithani V, Grant KA, McLauchlin J, dePinna E, Threlfall EJ. Assessment of the microbiological safety of dried spicesand herbs from production and retail premises in the United Kingdom.Food Microbiol. 2009;26(1):39–43.

36. Granum PE. Spotlight on Bacillus cereus and its food poisoning toxins. FEMSMicrobiol Lett. 2017;364(10):fnx071.

37. Yim JH, Kim KY, Chon JW, Kim DH, Kim HS, Choi DS, Choi IS, Seo KH.Incidence, antibiotic susceptibility, and toxin profiles of Bacillus cereus sensulato isolated from Korean fermented soybean products. J Food Sci. 2015;80(6):M1266–70.

38. Luna VA, King DS, Gulledge J, Cannons AC, Amuso PT, Cattani J.Susceptibility of Bacillus anthracis, Bacillus cereus, Bacillus mycoides, Bacilluspseudomycoides and Bacillus thuringiensis to 24 antimicrobials usingSensititre automated microbroth dilution and Etest agar gradient diffusionmethods. J Antimicrob Chemother. 2007;60(3):555–67.

39. Kim CW, Cho SH, Kang SH, Park YB, Yoon MH, Lee JB, No WS, Kim JB. Prevalence,genetic diversity, and antibiotic resistance of Bacillus cereus isolated from Koreanfermented soybean products. J Food Sci. 2015;80(1):M123–8.

40. Krause N, Moravek M, Dietrich R, Wehrle E, Slaghuis J, Martlbauer E.Performance characteristics of the Duopath(R) cereus enterotoxins assay forrapid detection of enterotoxinogenic Bacillus cereus strains. Int J FoodMicrobiol. 2010;144(2):322–6.

41. Danylec N, Gobl A, Stoll DA, Hetzer B, Kulling SE, Huch M. Rubneribacterbadeniensis gen. Nov., sp. nov. and Enteroscipio rubneri gen. Nov., sp. nov.,new members of the Eggerthellaceae isolated from human faeces. Int J SystEvol Microbiol. 2018;68(5):1533–40.

42. Bankevich A, Nurk S, Antipov D, Gurevich AA, Dvorkin M, Kulikov AS, LesinVM, Nikolenko SI, Pham S, Prjibelski AD, et al. SPAdes: a new genomeassembly algorithm and its applications to single-cell sequencing. J ComputBiol. 2012;19(5):455–77.

43. Wattam AR, Davis JJ, Assaf R, Boisvert S, Brettin T, Bun C, Conrad N, DietrichEM, Disz T, Gabbard JL, et al. Improvements to PATRIC, the all-bacterialbioinformatics database and analysis resource center. Nucleic Acids Res.2017;45(D1):D535–42.

44. Zankari E. Comparison of the web tools ARG-ANNOT and ResFinder fordetection of resistance genes in bacteria. Antimicrob Agents Chemother.2014;58(8):4986.

45. Carattoli A, Zankari E, Garcia-Fernandez A, Voldby Larsen M, Lund O, Villa L,Moller Aarestrup F, Hasman H. In silico detection and typing of plasmidsusing PlasmidFinder and plasmid multilocus sequence typing. AntimicrobAgents Chemother. 2014;58(7):3895–903.

Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.

Fiedler et al. BMC Microbiology (2019) 19:250 Page 13 of 13